Server hit count down since implementing CSS

Unanswered Question
Aug 6th, 2007

We have had a CSS in place for about two months now.

The CSS is configured in one armed mode with groups and "add destination service X" configured for our server.

Since we have implemented the CSS the hit count on the server has gone down from 200,000 hits a month to about 70,000.

The main server has not failed to the failover server in that time, other than a few minutes for testing several times amounting to a few hundred hits.

Could that fact the the hits are now originating from the CSS group address and not the actual IP address of the client affect the hit count in some way?

It seems that if that were the case, they would have dropped more than that.

But, if there is a hit counter, I am thinking it shoulnd't matter the source ip hitting the box.

Are there any thoughts?

I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 5 (1 ratings)
Loading.
Diego Vargas Mon, 08/06/2007 - 09:36

You mean hits on your server right? If so, what does your server count as a hit? a request, a packet or a new connection.

Is your CSS configure with a layer 5 content rule? If so, then the CSS will be answering the SYNs with a SYN/ACK before passing the request to the server, that might cause the server to log less hits.

wilson_1234_2 Mon, 08/06/2007 - 09:50

Thanks for the reply.

yes, I am talking about hits on the server. We have webtrends and have noticed that the hitcounts on the server is 20% less that it was before implementing the CSS.

Below is the content rule.

Forgive my questions:

How would answering the syn with syn/ack reduce the hits if the user still gets passed to the server?

Can you explain that process, are you talking about the keepalives?

With the http keepalive configured as below, would this be considered a hit on the server?

Could it be that the hitcount has increased to the point that the counter is starting overagain?

service MCI-xxxx.com

protocol tcp

port 80

keepalive type http

keepalive uri "/index.asp"

keepalive retryperiod 15

keepalive frequency 15

ip address 2.1.1.75

active

owner MCI-xxxx.com

content MCI-xxxx.com-http-rule

add service MCI-xxxx.com

balance aca

primarySorryServer MCI-xxxx.com-dr

secondarySorryServer MCI-xxxx.com-backupredirect

vip address 2.1.1.71

protocol tcp

port 80

url "/*"

active

Diego Vargas Mon, 08/06/2007 - 09:58

Hi,

Unfortunately I am not able to tell you what would you server consider as a HIT, actually from the CSS perspective I would like to know that to have an idea of what change on having the CSS could cause the server to log more or less hits.

You need to find what exactly would your server consider as a hit.

wilson_1234_2 Mon, 08/06/2007 - 10:04

Thanks for the reply.

My thought was, it shouldn't matter correct, if the user get to the webpage, it should be a hit coming from the CSS no matter what the source address is?

Also did you see my question about the keepalives above?

Diego Vargas Mon, 08/06/2007 - 10:18

How would answering the syn with syn/ack reduce the hits if the user still gets passed to the server?

That is why I am asking what would your server consider a hit, a packet? a request? a new connection?

With the http keepalive configured as below, would this be considered a hit on the server?

I do not know what the server would log as a hit, with this configuration, the CSS is sending a HTTP HEAD and waiting for a 200 Ok, I guess the server should consider a HEAD request as a hit.

Could it be that the hitcount has increased to the point that the counter is starting overagain?

This is information you need to get from the server people.

Actions

This Discussion